Why Enrico Fermi is the Father of Product Management

Enrico Fermi was an Italian-American physicist who helped build the world’s first nuclear reactor. That’s not why I love him though. I love him for the Fermi method.

The Fermi method is a prediction technique where you break down a problem into a bunch of steps and make educated guesses on each part.

His most famous example is asking “How many piano tuners are there in Chicago?” Let’s work through it together.

1. I live in Houston and I think Houston has 3M–4M people. I also know it’s predicted overtake Chicago soon as America’s 3rd largest city. So I’m 80% confident Chicago has 3.5M — 4.5M people. I’ll split the difference and say it has 4M.
2. Next, I have to figure out what % of people have a piano in their home. I know it’s an expensive item that not everyone can afford. And most people that can afford it, don’t want it. Maybe 1/50 households have a piano. And I’ll say there’s an average of 2.5 people in an household. So there’s 1 piano per 125 people.
3. Schools, concert halls, etc. have pianos too. Maybe that’s another 1 piano per 125 people.
4. I’d guess a piano tuner can tune 2 pianos per day once you factor in transportation time, bathroom breaks, lunch, etc.
5. There are 260 business days in a year, but he probably only works 240 (taking into account holidays, vacation days, and sick days).

So let’s put that all together.

1. There are 4M people in Chicago * (2 pianos / 125 people) = 64,000 pianos
2. A piano tuner can tune 2 pianos per day * 240 days = 480 pianos per year
3. 64,000 pianos / 480 pianos tuned per year = 133 piano tuners

What’s the real answer? According to an analysis by Wolfram Alpha, it’s 350. That’s pretty close!

And if you look closely, they were using the Chicago metro area while I was using the city proper. The Chicago metro area has ~9.5M people while the city proper has ~2.7M. If they just used the city proper, they would have estimated 99 piano tuners. I was pretty darn close!

The reason the Fermi method works so well is around half your estimates will be too high and half will be too low. And as such, they cancel each other out.

The Fermi method is the cornerstone of good product management and decision making. You should apply it to every feature you build.